Add R2 change for the interaction for moderation.

From Kelsey Sewell:

Update:

I tried converting Demographic.E4 allele to a factor with 2 levels (as opposed to a character variable) and I still was getting the same error. Please note, there are no NA’s in this variable:
summary(ImputesAmyloid$Demographic.E4 allele)
No Yes
173 51

I then tried to recreate the error using the states dataset from R and I couldn’t. I used this code and it ran fine:
states = as.data.frame(state.x77)
states$murdercat = states$Murder
states$murdercat= as.factor(ifelse(states$Murder <= 5, "Low",
ifelse(states$Murder > 5, "High", NA)))
test = moderation1(y = "Income",
x = "Illiteracy",
m = "murdercat",
cvs = "Frost",
df = states,
with_out = T)

I so I created a new ‘categorical’ variable in my dataset (even though Demographic.E4 allele was already categorical):
ImputesAmyloid$allelecat = ImputesAmyloid$Demographic.E4 allele
ImputesAmyloid$allelecat = as.factor(ifelse(ImputesAmyloid$Demographic.E4 allele == "No", "No",
ifelse(ImputesAmyloid$Demographic.E4 allele == "Yes", "Yes", NA)))

Then ran the model using the new variable and it ran fine.

doublemodtest <- moderation1(y = "Image.PET.Centiloid",
                  x = "totalMETs",
                  m = "allelecat",
                  cvs = c("Age", "Demographic.Sex", "mood.y", "premorbidIQ.y", "`PSQI.Q7-Medicine-Induced Sleep`", "Collection"),
                  with_out = F,
                  df = ImputesAmyloid)

I really have no idea why this worked as Demographic.E4 allele and allelecat appear to be exactly the same?

str(ImputesAmyloid$Demographic.E4 allele)
Factor w/ 2 levels "No","Yes": 1 1 1 1 1 2 1 1 1 1 ...
str(ImputesAmyloid$allelecat)
Factor w/ 2 levels "No","Yes": 1 1 1 1 1 2 1 1 1 1 ...

The only difference I can think of is that Demographic.E4 allele was read in from excel using the readxl package, whereas allelecat wasn’t.

I'm not sure if anyone has similar problems but I get this when I try to download MeMoBootR in R via

devtools::install_github("doomlab/MeMoBootR")

I get this message in the end (in the console):

Downloading GitHub repo doomlab/MeMoBootR@HEAD
These packages have more recent versions available.
It is recommended to update all of them.
Which would you like to update?

1: All
2: CRAN packages only
3: None
4: rlang (0.4.10 -> 0.4.11) [CRAN]
5: pillar (1.4.7 -> 1.6.1 ) [CRAN]
6: fansi (0.4.2 -> 0.5.0 ) [CRAN]
7: tibble (3.1.1 -> 3.1.2 ) [CRAN]
8: shape (1.4.5 -> 1.4.6 ) [CRAN]

Enter one or more numbers, or an empty line to skip updates:
Error: Failed to install 'MeMoBootR' from GitHub:
System command 'Rcmd.exe' failed, exit status: -1, stdout + stderr empty

This is when I don't update any of the packages. Updating the pacakges results in the same thing but with more errors, if that helps:

devtools::install_github("doomlab/MeMoBootR")
Downloading GitHub repo doomlab/MeMoBootR@HEAD
These packages have more recent versions available.
It is recommended to update all of them.
Which would you like to update?

1: All
2: CRAN packages only
3: None
4: rlang (0.4.10 -> 0.4.11) [CRAN]
5: pillar (1.4.7 -> 1.6.1 ) [CRAN]
6: fansi (0.4.2 -> 0.5.0 ) [CRAN]
7: tibble (3.1.1 -> 3.1.2 ) [CRAN]
8: shape (1.4.5 -> 1.4.6 ) [CRAN]

Enter one or more numbers, or an empty line to skip updates: 1
rlang (0.4.10 -> 0.4.11) [CRAN]
pillar (1.4.7 -> 1.6.1 ) [CRAN]
fansi (0.4.2 -> 0.5.0 ) [CRAN]
tibble (3.1.1 -> 3.1.2 ) [CRAN]
shape (1.4.5 -> 1.4.6 ) [CRAN]
Installing 5 packages: rlang, pillar, fansi, tibble, shape
Installing packages into ‘C:/Users/user/Documents/R/win-library/3.6’
(as ‘lib’ is unspecified)

There are binary versions available but the source versions are
later:
binary source needs_compilation
pillar 1.6.0 1.6.1 FALSE
fansi 0.4.2 0.5.0 TRUE
tibble 3.1.1 3.1.2 TRUE
shape 1.4.5 1.4.6 FALSE

trying URL 'https://cran.rstudio.com/bin/windows/contrib/3.6/rlang_0.4.11.zip'
Content type 'application/zip' length 1224139 bytes (1.2 MB)
downloaded 1.2 MB

package ‘rlang’ successfully unpacked and MD5 sums checked
Warning: cannot remove prior installation of package ‘rlang’
Warning: restored ‘rlang’

The downloaded binary packages are in
C:\Users\user\AppData\Local\Temp\Rtmpc9VAJT\downloaded_packages
installing the source packages ‘pillar’, ‘fansi’, ‘tibble’, ‘shape’

trying URL 'https://cran.rstudio.com/src/contrib/pillar_1.6.1.tar.gz'
Content type 'application/x-gzip' length 1111747 bytes (1.1 MB)
downloaded 1.1 MB

trying URL 'https://cran.rstudio.com/src/contrib/fansi_0.5.0.tar.gz'
Content type 'application/x-gzip' length 301724 bytes (294 KB)
downloaded 294 KB

trying URL 'https://cran.rstudio.com/src/contrib/tibble_3.1.2.tar.gz'
Content type 'application/x-gzip' length 686145 bytes (670 KB)
downloaded 670 KB

trying URL 'https://cran.rstudio.com/src/contrib/shape_1.4.6.tar.gz'
Content type 'application/x-gzip' length 677262 bytes (661 KB)
downloaded 661 KB

The downloaded source packages are in
‘C:\Users\user\AppData\Local\Temp\Rtmpc9VAJT\downloaded_packages’
Error: Failed to install 'MeMoBootR' from GitHub:
System command 'Rcmd.exe' failed, exit status: -1, stdout + stderr empty
In addition: Warning messages:
1: In file.copy(savedcopy, lib, recursive = TRUE) :
problem copying C:\Users\user\Documents\R\win-library\3.6\00LOCK\rlang\libs\x64\rlang.dll to C:\Users\user\Documents\R\win-library\3.6\rlang\libs\x64\rlang.dll: Permission denied
2: In i.p(...) : installation of package ‘fansi’ had non-zero exit status
3: In i.p(...) : installation of package ‘shape’ had non-zero exit status
4: In i.p(...) : installation of package ‘pillar’ had non-zero exit status
5: In i.p(...) : installation of package ‘tibble’ had non-zero exit status

Does anyone know how to deal with this? I tried running RStudio as admin and that didn't work.

Hi Erin!

It appears that the boot.ci function only outputs normal cis. I feel like there should be an argument to specify what type of cis you want to perform. Am I missing something? If not, can I make a pull request and change?

Thanks for these great tutorials. I just followed your youtube video for mediation1, but my output is different. for example, when showing "model$boot.results" i have 2 lines of output that give results at 2 different levels of cyl (cyl6 and cyl8). i don't think i turned on this function, since i made sure the model was identical to that in your script in the video... Is this an intentional change in the output of the mediation1?

Hi,
Thanks a lot for this package, and the youtube videos too.
I have a model where one variable is a mediator for a few independent variables and at the same time a moderator for a few other independent variables. I am not sure which example is the best to work on my data.
I appreciate if you could help me to know if the problem is solvable.

Regards,
Saeed

Hi and thanks for this great package,

I was wondering how one can calculate P for the indirect effect in the mediation2 func?

Best/Oveis

Dear Dr. Erin M. Buchanan,

Thank you very much for developing this R package and providing code, examples, and more! It is so educational and I appreciate it greatly.

My data suffers missing data (in the dv), and I think it is why get the following error:

Error in data.frame(..., check.names = FALSE) :
arguments imply differing number of rows: 122, 72
In addition: Warning messages:
1: In badmahal + badleverage :
longer object length is not a multiple of shorter object length
2: In badmahal + badleverage + badcooks :
longer object length is not a multiple of shorter object length

I did not manage to find an argument for excluding NAs; is that correct? Do you know if there is a way around this problem?

Furthermore, I wonder if the mediation1 function could be used for a categorical M? I have found the processR (catMediation)
package, but again no argument for omitting/excluding NAs.

Thank you very much for your time.

Best regards,
Emma

Hello!

First of all, thank you for this amazing package and all of your incredible YouTube videos. I have really learned a lot!

I was doing a mediation with a dichotomous X, dichotomous Y, and a continuous M using the mediation1() function. When it runs I get the error Error in mediation1(y = "SuicideBIN_Text", x = "cp_np", m = "rh_G_postcentral_thickness", : Y should not be a categorical variable. Log regression options are coming soon.

I was wondering if there was an update coming soon for this. Thanks for all your help!